Accurate quantification of multi-animal behavior plays a pivotal role in unraveling the intricacies of animal social interactions with far-reaching applications in neuroscience and ecology.
In a study published in Nature Machine Intelligence, researchers from the Brain Cognition and Brain Disease Institute (BCBDI) of the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences have proposed a few-shot learning artificial intelligence (AI) framework, the Social Behavior Atlas (SBeA), for multi-animal three-dimensional (3D) social pose estimation, identification, and behavior embedding.
The new technology can identify similar-looking animals with more than 90 percent accuracy, eliminating the need to define social behavior categories in advance, which may help discovery new, undefined differences in animal social behavior.
The new algorithm can also effectively synthesize a large volume of new data and train models with higher accuracy, thus obtaining more accurate 3D social gesture estimation results, said Wei Pengfei, correspondence author of the study.
SBeA is suitable for accurately calculating 3D social posture, identity and fine social modules in mice, birds and domestic dogs, with potential for cross-species applications, Wei said.
